This study examines how CEO power affects earnings quality and investigates the moderate influence of institutional ownership and the controlling shareholder on this relationship. The study sample includes 44 non-financial EGX-100 companies with 220 balanced observations covering 2017-2021. Four accounting-based measures of Francis et al. (2004) (accrual quality, earnings persistence, predictability, and income smoothing) are used as earnings quality proxies. Based on Finkelstein (1992), this study used three resources of CEO power: structural power, ownership power, and expert power. Multivariate linear regression analysis is applied to panel data using Feasible Generalized Least Squares (FGLS). Results show that CEO's structure and ownership power positively impact accrual quality, earnings persistence, and predictability. However, the CEO's expert power negatively impacts earnings persistence and predictability. The findings suggest that although institutional ownership and the controlling shareholder as individual variables negatively affect earnings quality indicators, they mitigate CEO expert power's negative impact and enhance the CEO power indicators' effects on earnings quality indicators. In addition, the z-score and corporate governance efficiency positively impact earnings persistence and predictability. This finding shows that firms with solid financial positions are more prone to have earnings quality. Further, corporate governance efficiency can prevent power abuses and ensure the CEO is in a firm's interest. Firm size positively influences earnings quality proxies, while financial leverage negatively affects them, supporting the political cost and debt covenant hypotheses in positive accounting theory.
Based on the existing literature, limited attention was paid to the association between CEO power and earnings quality, thus highlighting the novelty and significance of this study. Further, this study examines the potential moderating influence of the controlling shareholder and institutional ownership on such a relationship that has not been discussed in prior research.